TOPIC 2 STATES A NEED FOR PREDICTION OF MICROSTRUCTURE EVOLUTION PREDICTION OF SIZE F DEFECTS AND MODEL VALIDATION. TO ADDRESS THESE NEEDS WE PROPOSE TO USE MESOSCALE MODELING OF A) THE SOLIDIFICATION PROCESS INCLUDING TWO DIFFERENT POROSITY MECHANISMS AND B) THE SUBSEQUENT HEAT TREATMENTS. A PARTICULAR CHALLENGE IS TO MODEL THE DEVELOPMENT OF POROSITY IN THE PRINTING OF METAL PARTS BECAUSE OF THE WELL-KNOWN CONNECTION BETWEEN FATIGUE CRACK INITIATION AND THE UPPER TAILS IN DEFECT SIZE DISTRIBUTIONS. FOR IN 718 PARTS PRODUCED BY POWDER BED FUSION THE POTENTIAL SOURCES OF POROSITY ARE KEYHOLING LACK OF FUSION AND PORES INHERITED FROM THE POWDERS. KEYHOLING IS AVOIDED BY MAINTAINING A SUFFICIENTLY LOW POWER DENSITY AND WILL NOT BE CONSIDERED IN THIS PROJECT BECAUSE TYPICAL PRINTING CONDITIONS AVOID THIS PROBLEM. WITH REGARDS TO LACK OF FUSION POROSITY WE HAVE DEVELOPED SOFTWARE TOOLS TO PREDICT THE VOLUME PERCENTAGE SIZE AND SHAPE OF UNMELTED REGIONS IN THE PART. THE PREDICTIONS ALLOW LACK OF FUSION PORES TO BE EXCLUDED WHILE MAXIMIZING BUILD RATES AND AVOIDING LENGTHY EXPERIMENTAL TRIALS. THESE PREDICTIONS HAVE BEEN TESTED AGAINST LITERATURE RESULTS AND IN THIS PROJECT THE ACCURACY OF THE PREDICTIONS AT MAXIMIZED BUILD RATES WILL BE EVALUATED FOR IN 718. FOR PORES ARISING FROM THE POWDER PARTICLES THEMSELVES THE INITIAL HYPOTHESIS IS THAT PORES BUBBLES IN THE METAL POWDER PARTICLES PERSIST IN THE MELTING AND SOLIDIFICATION PROCESS. THIS IS POSSIBLE BECAUSE OF THE SHORT TIMES DURING WHICH A GIVEN VOLUME ELEMENT IS MOLTEN AND THE FINITE RISE RATES OF BUBBLES IN MELTS. WE PROPOSE TO USE THE HYBRID POTTS CELLULAR AUTOMATON MODEL BECAUSE IT IS EFFICIENT AND READILY ADAPTABLE TO DIFFERENT PROCESSES AS EVIDENT IN THE LITERATURE WHERE IT HAS FOR EXAMPLE BEEN APPLIED TO HIGH SPEED WELDING. THE HYPOTHESIS WILL BE TESTED BY INTRODUCING VOIDS INTO THE MELT POOLS AS THEY WOULD BE INHERITED FROM POWDER PARTICLES. VIA STANDARD SIMULATION METHODS THEY WILL BE ALLOWED TO MIGRATE AT RATES CORRESPONDING TO PHYSICAL BUOYANCY FORCES AT THE SAME TIME AS SOLIDIFICATION PROCEEDS. MATERIAL PROPERTIES WILL BE TAKEN FROM THE LITERATURE FOR THE IN 718 THAT IS THE MATERIAL OF GREATEST RELEVANCE TO NASA NEEDS. THE PROPOSED SIMULATION PROCEDURE WILL LEAD TO A PREDICTION OF MICROSTRUCTURE THAT INCLUDES THE RETAINED VOIDS AND PERMITS PROCESS VARIATIONS SUCH AS POWER AND TRAVEL SPEED TO BE EXPLORED. THREE POSSIBILITIES EXIST FOR VALIDATION. THE MOST OBVIOUS IS COMPARISON TO IN 718 POWDERS AND PRINTED METAL AND SUCH INFORMATION WILL BE AVAILABLE FROM EXISTING EFFORTS AT CMU AND VIA COLLABORATIONS WITH RELEVANT NASA GROUPS. THE SECOND IS FROM DETAILED SIMULATIONS OF MELTING FOR EXAMPLE FROM LAWRENCE LIVERMORE NATL. LAB. THE THIRD IS FROM HIGH SPEED X RAY RADIOGRAPHY FOR EXAMPLE WITH FRAME RATES APPROPRIATE TO THE MELT SOLIDIFICATION TIMES IN METAL PRINTING THAT IS UNDER DEVELOPMENT AT LIGHT SOURCES SUCH AS THE ADVANCED PHOTON SOURCE. EACH OF THESE SOURCES OFFERS DIFFERENT TYPES OF DATA AND THE AIM WILL BE TO OPTIMIZE THE SIMULATION AGAINST ALL AVAILABLE DATA. IN ADDITION TO THE GRAIN STRUCTURE AND DEFECT STRUCTURE THE PHASE COMPOSITION OF PRINTED MATERIAL WILL BE CALCULATED WITH STANDARD COMPUTATIONAL THERMODYNAMICS PACKAGE SUCH AS FACTSAGE OR THERMOCALC. A LONGER TERM OBJECTIVE OF THE DEVELOPMENT OF THE SIMULATION METHODOLOGY WILL BE TO ADDRESS THE LARGE STORED ENERGIES AND ORIENTATION GRADIENTS THAT ARE KNOWN TO OCCUR IN PRINTED METALS IN GENERAL AND IN 718 IN PARTICULAR. IN THIS CASE THE HYPOTHESIS IS THAT SOLIDIFICATION SHRINKAGE LEADS TO PLASTIC DEFORMATION DURING SUCCESSIVE PASSES OF THE MELT BEADS. SINCE CYCLIC LOADING FATIGUE IS OF PRIME INTEREST THE MAIN FOCUS OF THE EFFORT WILL BE ON DEVELOPING A PREDICTIVE CAPABILITY FOR POROSITY AS A FUNCTION OF THE PRINTING PARAMETERS.